Journal of Geophysical Research: Space Physics

Observations of DC electric fields in the low-latitude ionosphere and their variations with local time, longitude, and plasma density during extreme solar minimum

Authors


Abstract

[1] DC electric fields and associated E × B plasma drifts detected with the double-probe experiment on the C/NOFS satellite during extreme solar minimum conditions near the June 2008 solstice are shown to be highly variable, with weak to moderate ambient amplitudes of ∼1–2 mV/m (∼25–50 m/s). Average field or drift patterns show similarities to those reported for more active solar conditions, i.e., eastward and outward during day and westward and inward at night. However, these patterns vary significantly with longitude and are not always present. Daytime vertical drifts near the magnetic equator are largest in the prenoon sector. Observations of weak to nonexistent prereversal enhancements in the vertical drifts near sunset are attributable to reduced dynamo activity during solar minimum as well as seasonal effects. Enhanced meridional drifts are observed near sunrise in certain longitude regions, precisely where the enhanced eastward flow that persisted from earlier local times terminates. The nightside ionosphere is characterized by larger-amplitude, structured electric fields dominated by horizontal scales of 500–1500 km even where local plasma densities appear relatively undisturbed. Data acquired during successive orbits indicate that plasma drifts and densities are persistently organized by longitude. The high duty cycle of the C/NOFS observations and its unique orbit promise to expose new physics of the low-latitude ionosphere.

1. Introduction

[2] This paper presents initial results from the Vector Electric Field Investigation (VEFI) on the Communication/Navigation Outage Forecast System (C/NOFS) satellite. The DC electric fields and associated E × B plasma drifts presented here were acquired between 9 May 2008 and 4 July 2008. The selected period covers a significant fraction of a full precessional period of the C/NOFS orbit and allows sampling at all local times and longitudes within the altitude and inclination limits of the C/NOFS satellite orbit (described below). The data presented here were gathered under nearly homogeneous solar minimum and solstice conditions. Within this limited context, we demonstrate how the DC electric fields and their associated E × B drifts varied as functions of local time, longitude, and plasma density.

[3] Many previous studies have demonstrated the controlling influence of electric currents and fields on the dynamics of the low-latitude ionosphere [e.g., Rishbeth, 1971; Fejer, 1997; Heelis, 2004]. Driving forces include (1) ever-present gravity-driven electric currents [e.g., Eccles, 2004], (2) the solar quiet (Sq) current system driven by the dayside E region dynamo process [e.g., Farley et al., 1986; Eccles, 1998a, 1998b], (3) the wind-driven F region dynamo which produces strongest electric fields at night [Rishbeth, 1971], (4) tides and planetary waves originating in the troposphere and stratosphere [e.g., Immel et al., 2006; Liu et al., 2010], and (5) magnetic storm related contributions. The storm contributions include dusk-to-dawn disturbance dynamo fields driven by winds excited by auroral energy [e.g., Blanc and Richmond, 1980; Fejer and Scherliess, 1995, 1997; Scherliess and Fejer, 1997], dawn-to-dusk fields penetrating from high latitudes [e.g., Nopper and Carovillano, 1978], and dusk-to-dawn, overshielding fields during periods of Region 2 dominance early in recovery phases of storms and substorms [e.g., Kikuchi et al., 2000, 2008]. The interaction of these currents and fields with conductance gradients leads to the formation of polarization electric fields that are generated internally to insure that the ionospheric currents remain divergence free. In some cases, the local polarization fields can be much stronger than the vector sum of the imposed large-scale fields.

[4] Vertical drifts caused by the eastward electric field components are characterized by prereversal enhancements (PREs) near the dusk terminator. The PRE is a polarization effect required to maintain the continuity of Sq currents near the dusk meridian [e.g., Farley et al., 1986; Eccles, 1998b]. At postsunset local times, plasma initially rises in response to negative polarization charges that accumulate near the terminator [Eccles, 1998a]. After ∼2000 LT the vertical drift usually reverses thereby stabilizing the bottomside of the F layer against irregularity growth. Ezonal normally remains westward across the nightside. The intensities of PRE vertical drifts show seasonal, longitudinal, and solar cycle dependencies, tending to be largest near equinoxes at solar maximum [Scherliess and Fejer, 1997]. During the magnetically quiet times of solar minimum, PRE drifts become very weak [Fejer and Scherliess, 1995].

[5] The presence of eastward Ezonal in the immediate postsunset hours renders the bottomside of the equatorial ionosphere unstable to the growth of the generalized Rayleigh-Taylor instability [e.g., Scannapieco and Ossakow, 1976; Ott, 1978]. The PRE opens a finite window of opportunity for equatorial plasma depletions to form and grow in the equatorial ionosphere before the damping action of westward Ezonal becomes dominant. At satellite altitudes in the topside ionosphere depletions usually appear as upward moving flux tubes of low plasma density. Composition analysis showed that the depleted flux tubes observed above the F peak originated at bottomside altitudes [Hanson and Sanatani, 1973]. The rapid upward motion of depleted flux tubes is due to eastward, small-scale polarization electric fields [e.g., Aggson et al., 1992]. Recently, examples of upward moving plasma density enhancements have also been reported [Le et al., 2003; Park et al., 2008].

[6] Previous in situ measurements, notably on the Atmospheric Explorer-E, Dynamics Explorer-2, San Marco, DMSP, and ROCSAT satellites have provided direct measurements of low-latitude electric fields and/or plasma drifts [see, e.g., Coley and Heelis, 1989; Maynard et al., 1988; Fejer et al., 1995; Maynard et al., 1995; Hartman and Heelis, 2007; Fejer et al., 2008]. Furthermore, detailed observations of plasma drifts as a function of altitude and local time at a given longitude are routinely provided by incoherent scatter observations, most notably by the Jicamarca radar [e.g., Fejer et al., 1979, 1991; Scherliess and Fejer, 1997]. The systematics of background electric fields at low magnetic latitudes have thus been inferred from synoptic averages of large-scale, in situ and radar databases gathered by these measurement techniques under a variety of seasonal and solar conditions.

[7] The San Marco mission provided the only other direct measurements of ionospheric DC electric fields by sensors on an equatorial orbiting (i.e., low inclination) satellite, and these previous measurements are thus highly germane to the data presented here. The spacecraft was launched into a 2.9° inclined orbit in 1988, a period of moderate solar activity. The double-probe electric field instrument utilized two pairs of wire antennas in the spin plane with 35 m separation between the sensors. Owing to severe power constraints, San Marco had a limited duty cycle, with data typically being acquired for only a small fraction of each orbit. Nevertheless, the San Marco electric field observations yielded instantaneous zonal and meridional electric fields that were invaluable for both case studies and statistically averaged data that were presented as E × B drifts, enabling longitudinal effects to be discerned [Maynard et al., 1995].

[8] The remainder of this paper is divided into four sections. Section 2 provides brief descriptions of the VEFI and Planar Langmuir Probe (PLP) sensors on C/NOFS. We then present the observations in terms of illustrative examples and synoptic/statistical summaries of electric field and plasma drifts acquired between May and early July 2008. Sections 4 and 5 relate C/NOFS observations to previously reported measurements and discuss the data in terms of an emerging new picture of ionosphere-thermosphere coupling during the prevailing extreme solar minimum conditions.

2. Experiment Description

[9] The C/NOFS mission is designed to study and predict scintillations in the equatorial ionosphere [de La Beaujardière et al., 2004]. Launched on 17 April 2008 into a 401 km by 867 km orbit with a 13° inclination, C/NOFS includes a suite of instruments to measure ionospheric and neutral parameters. Those germane to this study include the Vector Electric Field Investigation (VEFI) that provides electric field and magnetic field observations and the Planar Langmuir Probe (PLP) that measures plasma densities and temperatures. The satellite includes three-axis attitude control and body-mounted solar panels to avoid disrupting the electric field measurements. A star sensor provides attitude knowledge to better than 0.1°.

[10] Vector measurements of the DC electric field on the C/NOFS satellite are provided by a three-axis detector that gathers DC and AC (or wave) electric fields using the double probe technique [e.g., Maynard, 1998]. The instrument includes six rigid 9.5 m booms that extend 11.8 cm diameter spherical sensors with embedded preamplifiers. The booms are oriented to provide three orthogonal, 20 m tip-to-tip double probes that detect the three components of the vector DC and AC electric fields. DC electric fields are sampled at 1024 times per second by digitizing the potential differences between opposing sensors using 16-bit A/D converters. The standard DC electric field waveforms are subsequently filtered and averaged on board the satellite to produce three, 16 sample/sec components prior to telemetry to the ground. Advanced processing techniques are used to identify and subtract contact potentials, offsets, and sheath-related electric fields in order to assure high quantitative accuracy [Maynard, 1998].

[11] Vector DC magnetic field data are also gathered as part of the VEFI instrument with a fluxgate magnetometer sensor that is located on a 0.6 m boom and includes active thermal compensation. The three orthogonal magnetic field components are detected within a ±45,000 nT dynamic range and are digitized at 1024 sample/sec and then averaged on board to 1 sample/sec, with the same 16-bit A/D converters that sample the DC electric field data. The simultaneous measurement of the ambient, vector magnetic field at the precise spacecraft location enables (1) identification and subtraction of −VSC × B electric fields, (2) computation of associated E × B plasma drift velocities, and (3) rotation of the electric field and plasma drift components to a reference frame defined by the ambient magnetic field.

[12] The Planar Langmuir Probe (PLP) was designed primarily to measure absolute ion densities and their fluctuations in the F region of the equatorial ionosphere. PLP sensors are conceptually similar to swept bias Langmuir probes and retarding potential analyzers that have flown on numerous satellite missions [e.g., Hanson and Heelis, 1975], albeit with advanced electronic capabilities. The instrument contains two independent sensors. The Surface Probe is an electrically isolated flat plate with controllable bias that is directly exposed to the space plasma environment. It is normally held at a potential that is negative with respect to spacecraft ground and thus measures saturation ion currents. The Ion Trap is a grounded electrode mounted beneath a stack of three 80% transmission etched gold grids. The outermost grid is held at, or near, spacecraft ground. The second grid is normally at −5V to allow passage of all ions, but can be swept from −5 to +12V to perform retarding potential analyses. The third grid is an electron repeller and photoelectron suppressor held at −12V. The Ion Trap current is initially sampled at 2048 sample/sec, passed through the appropriate anti-alias filters and then averaged down to the desired telemetry rate. Available sample rates are 32, 256, 512, and 1024 Hz. PLP is nominally operated at 32 Hz during the day and at 512 Hz during eclipse to reduce telemetry volume. Data presented in this paper were collected with the Ion Trap sensor.

3. Data Presentation

[13] Electric field measurements, averaged to 1 sample/s, are shown for four orbits in Figures 1a1d where the time series presentation is centered on local midnight in each panel. The second and third panels from the top show the meridional and zonal components of the DC electric fields, whereas the uppermost panel shows “arrow” plots of their associated E × B velocities. In this presentation, the drift vectors have been averaged by a factor of 20 both for clarity in the arrow-type display and to emphasize the larger-scale variations. The four orbits shown in Figures 1a1d were chosen as representative of data when perigee was located near midnight, near noon, near dusk, and near dawn. The coordinate system is such that the zonal direction is defined by B × R, where B the local magnetic field vector and R is the vector from the center of the earth to the spacecraft. The meridional component is the zonal direction × B and is positive outward. Note that at the magnetic equator, the meridional component is in the vertical direction.

Figure 1.

Examples of vector electric fields, associated E × B plasma drifts, and plasma number density gathered on four orbits of the C/NOFS satellite.

Figure 1.

(continued)

[14] The presented electric fields utilize average daily offsets for each orthogonal double-probe component that have been derived by assuming there are no DC electric fields along the magnetic field line (i.e., E · B = 0). This assumption is well established for electric field measurements in the ionosphere, particularly those gathered away from the high-latitude region, and has been validated using detailed comparisons of double-probe and ion drift measurements on previous satellites [Hanson et al., 1993]. Furthermore, the offsets derived for this experiment have been independently verified using occasional spins of the C/NOFS spacecraft carried out specifically to ensure the accuracy of the double-probe measurement. We also include in our analysis the assumption that the integrated zonal electric field is zero about a single orbit (i.e., Curl Ezonal = 0). Despite these procedures, small residual offsets may remain in the data, for example that pertain to asymmetric electron distributions around the spacecraft. Such nongeophysical contributions would not be expected to significantly alter the data or conclusions presented in this paper.

[15] The fourth panel from the top shows plasma densities at 1 sample/sec gathered from the PLP. Below this panel, a black bar delineates times when the spacecraft was in the Earth's shadow. The bottom panel provides a map of the Earth's low-latitude regions with satellite locations superposed. Shaded segments mark nighttime portions of orbits with respect to the ground. Red and blue lines indicate satellite altitudes and magnetic equator locations, respectively.

[16] A number of features are immediately evident in the electric field data shown in Figures 1a1d. We first note that in each example, plotted electric fields varied smoothly on the dayside but were highly structured across the nightside. Notice further that the corresponding E × B drifts appear moderately organized, generally changing direction between day and night. Ambient E × B drifts observed on a given C/NOFS orbit, as indicated in Figures 1a1d, often follow established patterns of low-latitude ionospheric drifts: outward and westward during the day; inward and eastward at night. Although it is tempting to compare these drift measurements with global plasma circulation patterns, care must be exercised when interpreting drifts observed during a given orbit within the larger-scale, global electrodynamic picture. Satellites such as C/NOFS in low-inclination orbit continuously cross different longitude sectors where the ambient electric fields (drifts) are known to vary [e.g., Maynard et al., 1995; Fejer et al., 2008]. For example, in Figure 1b, observed daytime meridional drifts a few hours after local noon were inward. While this is not atypical of many of the afternoon meridional drifts observed thus far by C/NOFS, it is opposite to the direction of the more “conventional” post noon outward drift as well as the direction of the prereversal enhancement. Furthermore, the satellite samples symmetrically about the geographic rather than the magnetic equator. Since the offset between the geographic and magnetic equators varies with longitude, the spacecraft samples different ranges of magnetic latitude along its orbit as well as from one orbit to the next. Hence, electric fields encountered along the C/NOFS trajectory typically map along magnetic field lines to considerably different altitudes at the magnetic equator depending on the latitude and longitude where they were observed.

[17] Plasma densities are also more structured on the nightside than the dayside. Attention is directed to large amplitude density depletions that characterize the postmidnight sector in Figure 1a. These depletions correspond to equatorial plasma depletions whose associated perturbation electric fields, directed to the east and slightly outward, are superposed on ambient, slowly changing background fields. The perturbations are polarization electric fields that drive upward and westward plasma drifts, characteristic of typical motions of the depleted flux tubes [e.g., Aggson et al., 1992]. Somewhat surprising in this example is the presence of low-amplitude perturbation electric fields at postsunset and dawn local times away from the large plasma depletion structures.

[18] Such structured nighttime electric fields are routinely observed in the C/NOFS data examined thus far. In fact, it is not unusual for the DC electric fields in the nighttime ionosphere observed on C/NOFS to be quite structured even when plasma densities only manifest small to modest perturbation levels. Data plotted in Figures 1b and 1c illustrate this point, showing electric fields that exhibit a high degree of structure throughout the nightside, with corresponding sharp reversals in both the fields and the resulting meridional and zonal plasma flows. Note that strong plasma depletion activity is not associated with these electric field structures. Although their waveforms are difficult to discern on a logarithmic scale, small-amplitude density variations were present in these examples.

[19] The nightside electric field structure observed in the C/NOFS data is, in general, dominated by horizontal scale lengths on the order of 500–1500 km or more along the satellite orbit (corresponding to roughly 60–200 s in the satellite reference frame), and may be associated with large-scale neutral atmosphere waves and structures. The electric field data in Figure 1c exhibit a large-scale quasi-oscillatory waveform that corresponds to scale lengths of ∼1000 km along the satellite path. (The horizontal satellite velocity is approximately 7.5 km/sec.) These structures persist not only throughout the nightside and into the dawn portion of the orbit but also over a large range of satellite altitudes, which changed from ∼500 km to ∼800 km in this example.

[20] Figure 1d reveals an isolated, enhanced zonal electric field at sunrise that corresponds to a region of meridional upward drifts that are observed precisely where the eastward zonal drift ends, as shown in the eastward to upward “twist” of the E × B velocities in the uppermost panel. This feature is similar to San Marco electric field observations of enhanced, sunrise meridional drifts reported by Aggson et al. [1995] and will be discussed below.

[21] The nearly continuous data coverage provided by C/NOFS reveals how the electric fields and plasma densities vary from orbit to orbit as well as from day to day. This is illustrated in Figures 2a and 2b, which show data acquired during 75 consecutive orbits in each of two 5 day periods, namely 6–10 June 2008 for Figure 2a and 16–20 June 2008 for Figure 2b. In this presentation, plasma drifts and densities sampled during each orbit have been averaged to a cadence of 1 sample/minute and have been essentially “turned on their sides.” Orbits are arranged in vertical patterns with respect to day of year (or orbit number) as the abscissa and local time the ordinate. The top panels of Figures 2a and 2b show the strength of the meridional component of E × B drift (corresponding to the zonal electric field) represented by shades of blue (red) for outward (inward) motion. Note that for data gathered on a low-inclination satellite such as C/NOFS, plasma drifts in the magnetic meridional direction are oriented near, but not precisely along, the vertical direction unless the data are sampled at the magnetic equator. The second panel shows the zonal component of E × B drifts (corresponding to meridional electric fields) using the same color scheme and with positive corresponding to eastward drifts. In some cases, particularly at night, instrumental glitches preclude accurate electric field measurements. Grey pixels mark such regions. A faint line near 20 LT appears as a relatively sharp change in color gradation in the meridional drift panels which occurs at the dusk terminator (i.e., as the spacecraft enters eclipse) and is believed due to photoelectron or sheath effects and is not geophysical. The bottom panel of each plot shows plasma densities using a logarithmic color scale. For reference, dotted lines mark crossings of the prime meridian, 0° geographic longitude, in each panel. Over the sampled 5 day interval, the satellite apogee and perigee occurred at average local times of 12.7 h and 0.8 h, respectively, for the data shown in Figure 2a, and at local times of 16.6 h and 4.5 h, respectively, for the data shown in Figure 2b.

Figure 2.

Plasma drift components (derived from the electric field data) and plasma density for 75 consecutive orbits within two 5 day intervals of C/NOFS observations. See text for details.

[22] Plasma drifts and densities presented in Figures 2a and 2b show variations that reflect latitude, longitude, local time, and altitude sampling effects. Consider first plasma densities plotted in the lowest panel of Figures 2a and 2b. The dayside plasma density variations show patterns of “slanted streaks” in this presentation format which indicate that the density variations are primarily organized by longitude. The fact that the streaks extend over a large portion of the daytime orbits indicates that this organization is present over essentially the entire range of altitudes sampled by the C/NOFS satellite, including regions near apogee (i.e., >800 km). Though somewhat less distinct, this general alignment of the density variations with longitude is also visible in the nightside data, particularly in cases where the density variation streaks appear as minima or depressions between midnight and dawn, as shown in Figure 2b. To help understand this ordering of the data by longitude, consider that as the satellite moves progressively through consecutive orbits, it samples the same longitude region ∼90 min later in local time each orbit. Thus, a geophysical feature at a given longitude which changes slowly with universal time would appear as a slanted streak in the orbit alongside orbit, local time versus day of year format of Figures 2a and 2b. That density enhancements and reductions organize by longitude is confirmed by the fact that the slanted streaks align parallel to the dotted lines corresponding to the loci of a given longitude (in this case, chosen to be 0° for the purposes of illustration.)

[23] We next consider the distributions of E × B drifts in the top two panels of Figures 2a and 2b. Here, the plasma flow amplitudes and directions are also strongly organized in both local time and longitude. Typically, plasma drifts in the low-latitude ionosphere are characterized as outward and westward during the day and inward and eastward at night [e.g., Kelley, 1989], which are generally observed in the C/NOFS data. Notice, however, that the dayside meridional drift magnitudes in the top panels of Figures 2a and 2b reveal variations that are organized in streak patterns with slopes that mimic those of constant longitude, as discussed for the plasma density data above. For example, the data selected for the examples in Figures 2a and 2b show a preponderance of larger amplitude, positive meridional daytime drifts occurring at the longitude sector that maximizes near 0°. For longitudes near 180° however, data in these examples show much weaker meridional flows. Note also the general lack of pronounced “prereversal” enhancements in the outward plasma flow near sunset, a feature that often characterizes the electrodynamics of the equatorial ionosphere [e.g., Kelley, 1989]. We return to consider implications of this observation in section 4 below.

[24] In contrast to the meridional daytime drifts, zonal drifts on the dayside in the center panels of Figures 2a and 2b show variations between orbits that are much more subtle, with the suggestion of complex variations with longitude. Similar to the meridional drifts, zonal flows abruptly change direction in the late afternoon local time sector, revealing concentrated eastward flows that are particularly strong near sunset in Figure 2a. Between local midnight and dawn, typical variations in the zonal plasma drifts with longitude are somewhat difficult to discern because of the electric field structuring noted above with regard to Figures 1a1d as well as due the fact that the background fields were not as strong. Changes in zonal drifts observed when comparing Figures 2a and 2b may have been, in part, due to the different altitudes in which the data were sampled, as well as from source locations at different off-equatorial latitudes where the electric fields would then map along magnetic field lines to altitudes encountered by the C/NOFS satellite. Our next step is to compute the average drifts over a much larger time period to establish general flow patterns at the magnetic equator as a function of local time.

[25] Figure 3 shows average E × B plasma drifts and their standard deviations, gathered within ±5° of the magnetic equator between 9 May and 4 July 2008 and displayed as functions of local time. Data are included for only periods of low magnetic activity (Kp ≤ 3). Nighttime intervals with significant plasma depletion activity were also excluded from the average data. The averaged vertical drifts (top panel, Figure 3), were strongly upward only between ∼0600 LT to ∼1400 LT, with average amplitudes varying, in general, between roughly 20–35 m/s. Averaged meridional drifts decreased considerably in the afternoon, becoming near zero or slightly negative (downward) between 1400 LT and sunset, with little or no evidence of a prereversal enhancement near sunset. In the nighttime sector, the averaged vertical drifts were downward, with values of about 20 m/s.

Figure 3.

Average E × B drifts and their standard deviations within 5° of the magnetic equator as a function of local time for the interval of 9 May through 4 July 2008.

[26] Zonal drifts plotted in the lower panel of Figure 3 indicate average values of 5–15 m/s from 0700–0900 LT increasing to ∼25 m/s westward across the midday–early afternoon (1000–1600 LT) sector. The averaged zonal drift reversed flow direction to eastward in the late afternoon (∼1630 LT) sector and increased to ∼50 m/s eastward near 1930 LT before decreasing to values near ∼10–20 m/s after midnight. Average zonal drifts showed a suggestion of organized, eastward flows for an interval of a few hours before sunrise, from 0430 to 0630 LT, before returning to the near-zero morning flow speeds mentioned above. It is important to note that average values of nighttime flows are difficult to assess because of electric field structuring that is pervasive across the darkened equatorial ionosphere. This may contribute to relatively large standard deviations associated with the plasma drifts on the nightside compared to the dayside. On the other hand, larger standard deviations in both components also reflect the fact that all longitudes were bunched together in the averaging process presented here.

[27] Data shown in Figure 4 illustrate how average drifts near the magnetic equator vary with longitude. Average plasma drifts sampled within ±5° of the magnetic equator during the same May–July 2008 interval as Figure 3, are now sorted into eight, evenly spaced longitude bins each of 45° width and plotted in the vector format displayed in the top panels in Figure 1. Although the general trends are similar to those of the averaged drift data shown in Figure 3, differences between longitude sectors are clearly evident. In all longitude sectors, the plasma flow is generally upward on the morningside and downward in the late afternoon and evening. The largest average vertical drifts in this study were observed in the 0700–1000 LT sector at all longitudes. However, vertical E × B drifts in the 225° and 270° sectors show a larger shear near the dawn meridian than elsewhere, with distinct upward directed flows within narrow local time intervals near dawn and within these longitudes bins only. The 270° sector includes the longitude of the Jicamarca radar, 284° E. At local times after noon, drifts are weaker than in the postsunrise to noon sector, and can either be upward, near zero, or erratic. Furthermore, no longitude sector manifests increased vertical drifts near sunset (i.e., near 19 LT) such as those typically associated with a prereversal enhancement, except a very slight increase in the 0° sector.

Figure 4.

Average measured E × B drifts within 5° of the magnetic equator separated within eight equally spaced longitude bins for the interval of 9 May through 4 July 2008.

4. Discussion

[28] Electric field data presented for consideration in section 3 evolved on both large and small spatial scales. The large-scale fields generally represent continual global responses to thermospheric wind dynamos operating in the E and F regions at low latitudes (see reviews by Eccles [1998a] and Heelis [2004]) and tidal and planetary wave forcing from the stratosphere and troposphere during both daytime and nighttime conditions [e.g., Immel et al., 2006; Liu et al., 2010]. Electric fields at low latitudes also may have contributions associated with magnetic storms, such as those that arise from the disturbance dynamo and auroral oval [Blanc and Richmond, 1980] as well as from partial effects of penetration fields from high latitudes [Nopper and Carovillano, 1978] and/or overshielding by Region 2 currents [Kikuchi et al., 2000, 2008]. Smaller-scale, localized electric fields reflect the build up of polarization charges along plasma gradients, such as at the terminators and those associated with neutral structures/gravity waves and plasma density depletions.

[29] In general, meridional and zonal plasma drifts associated with large-scale electric fields reported here show more diurnal variability than similar, previously reported measurements. For example, average zonal E × B drift patterns inferred from electric fields measured by the double-probe instrument on the San Marco satellite were considerably larger and more organized. Maynard et al. [1995, Figure 3] indicates diurnal flow patterns that were consistently eastward and downward at night, and westward and upward on the dayside. Similar average vertical drift patterns were reported based on observations from the Atmospheric Explorer-E [Fejer et al., 1995] and ROCSAT [Fejer et al., 2008] satellites, and for both the zonal and meridional drift components observed on the Dynamics Explorer-2 satellite [Coley and Heelis, 1989]. Maynard et al. [1988] reported average zonal drifts from the electric field detector on Dynamics Explorer-2. Average vertical and zonal components from the Jicamarca radar were reported by Fejer et al. [1991]. However, these reported in situ examples all reflect responses under moderate to high solar UV forcing conditions.

[30] Conversely, data presented here were acquired under extreme solar minimum conditions. The average F10.7 for the May–July 2008 period was ∼70 solar flux units (1 sfu = 1022 W/m2-Hz). These solar conditions are significantly lower than those prevailing at the acquisition times of the cited ionospheric electric field and plasma drift measurements. Fejer et al. [1991] presented radar measurements from the Jicamarca Radio Observatory that were representative of low solar activity (F10.7 < 100 sfu) during the May–June 1974–1977 solar minimum period. Their reported vertical drifts were comparable to those observed when C/NOFS was at longitudes near the Jicamarca radar (284° E). Unpublished measurements from the ROCSAT spacecraft, taken near the magnetic equator with F10.7 < 100 sfu show similarly diminished vertical plasma drifts during the May–June interval (B. Fejer, personal communication, 2009).

[31] A significant consequence of the extremely low solar activity encountered by the C/NOFS satellite is that the ionosphere and thermosphere were considerably contracted in altitude [Heelis et al., 2009]. This implies that the neutral winds responsible for the F region “dynamo” operate at lower altitudes, which during this extremely low period of solar activity would be primarily below the perigee (401 km) of the C/NOFS satellite. Since electric fields driven by the F region dynamo at off-equatorial latitudes might still map along magnetic fields lines to altitudes encountered by the C/NOFS satellite, the lowering of the region where the F region dynamo is operating cannot, by itself, explain the weaker observed electric fields. However, since the F region dynamo contributes to the “standard” electric field and plasma drift patterns in the low-latitude ionosphere [e.g., Eccles, 1998a; Heelis, 2004] and since the electric fields observed by VEFI are weaker and more variable than previously reported measurements, it is reasonable to assume that the intrinsic F region dynamo is weaker during the solar minimum, June solstice period reported here.

[32] An important feature of the C/NOFS data reported here is the lack of a well-defined, prereversal enhancement (PRE) in the vertical drift data. In fact, some C/NOFS observations indicate plasma flows that in the afternoon sector can be downward rather than upward. C/NOFS observations suggest that the strong neutral winds believed to create the prereversal enhancement were themselves reduced. Other studies using radar and ionosonde data show that the PRE diminishes at all altitudes during periods of low solar activity, consistent with an F region dynamo explanation for this phenomena [Eccles, 1998a, 1998b]. Seasonal effects are also important. Fejer et al. [1995] reported the absence of prereversal enhancement in the AE-E satellite average data gathered in the May–August period for low solar flux conditions. Fejer et al. [2008] reported reduced prereversal enhancement signatures in the ROCSAT data near the June solstice even for moderate levels of solar activity.

[33] Nightside plasma drifts are also consistent with general expectations for observations during very low solar activity. The nighttime C/NOFS zonal drifts reported here are considerably smaller than those previously described for periods of larger solar activity. For example, Maynard et al. [1995] show nighttime, eastward zonal drifts, averaged for all seasons, of ∼160 m/s prior to midnight, reducing to ∼115 m/s after midnight, and daytime westward zonal drifts of ∼60 m/s. If we consider that the nighttime zonal drifts represent, to first order, an analog for the zonal neutral winds [e.g., Heelis, 2004], we speculate that the extreme solar minimum conditions including the resulting contracted ionosphere/thermosphere system, may result in significantly lower zonal wind amplitudes. This would be consistent with the considerably weaker nighttime zonal drifts observed by C/NOFS, particularly those after local midnight.

[34] C/NOFS data also support investigations of the longitude dependence of variability in the electric field's meridional and zonal components. Variations with longitude appear to be somewhat more pronounced for the meridional than for the zonal flows, as shown in Figures 2 and 4. One of the more prominent effects is the diminished vertical drifts in the afternoon sectors that for some longitudes are downward (e.g., 90°). Vertical drift behavior as a function of longitude as observed by VEFI is, in general, consistent with that detected by the ROCSAT satellite [Fejer et al., 2008], albeit at moderate solar flux conditions. Hartman and Heelis [2007] showed that average vertical drifts observed by DMSP at 840 km near 0930 LT near the magnetic equator also manifest strong dependencies on longitude and season including some longitude sectors that displayed average vertical drifts near zero or even negative. Similarly, in the San Marco electric field data near the June solstice, significant differences in average plasma drifts occurred between longitudes where the magnetic equator is north of the geographic equator (Indian sector) and where the converse relation applies (Peruvian sector).

[35] A significant feature of the longitude study presented in Figure 4 is the increased vertical drift just after sunrise within the 225° and 270° longitude sector bins. This feature was also reported in the ROCSAT vertical drifts by Fejer et al. [2008] for the same season (May–August) and same general longitude region (see data at −120° on the right side of Figure 3 of Fejer et al. [2008] as well as Figure 4 of this same article). Furthermore, using San Marco data, Aggson et al. [1995] reported enhanced zonal electric fields near sunrise in the June–early September season confined to this same longitude region. These authors explained the increased field as a consequence of the opposite signs of the sunrise terminator and the magnetic declination, resulting in differences in the conjugate E region densities and subsequent polarization electric field, in an analog to the PRE at sunset. Our longitudinal sort in Figure 4 shows that the range of enhanced vertical drift at sunrise covers the region where the magnetic field declination is eastward, while the terminator slant is westward.

[36] The fact that enhanced vertical drifts are confined in local time (near sunrise) and longitude is also shown in the data presented in the upper panel of Figure 2b, where such isolated increases in meridional drifts are highlighted by the circles near 6 LT on successive days from 17 to 20 June 2008. Data shown includes the electric fields from the orbit shown in Figure 1d as well as the data shown by de La Beaujardière et al. [2009] where the enhanced vertical fields were observed colocated with a deep plasma trough. Since the density data in the lowest panel of Figure 2b show troughs and enhancements organized by longitude throughout the nightside, this suggests that such large-scale density patterns are governed by planetary-scale processes such as tides and not primarily by local electric fields.

[37] Another important aspect of this phenomenon lies in the enhanced eastward zonal flow that begins at earlier local times within the predawn hours but ends abruptly where the enhanced meridional flow is observed. The relation of the zonal and meridional drifts is shown explicitly in the data near sunrise (1430 U.T.) in Figure 1d. Here, the zonal flow slightly increases and then ends precisely where the outward flow enhancement is observed, maintaining flow continuity at least in the plane perpendicular to the magnetic field. Away from the localized region of the enhanced meridional outward flow, the relation of the eastward zonal flow and the enhanced meridional drift is difficult to discern because the satellite cannot sample all local times and longitudes at the same time. In other words, when examining geophysical features that are confined in local time and longitude, it is difficult to examine neighboring local times at the same longitude simply due to the limitations imposed by the satellite orbit.

[38] Another aspect of data shown in Figures 2a and 2b is the relationship of the large-scale plasma drifts and the corresponding variability of the simultaneously measured plasma densities. Enhancements in the drifts, particularly the meridional components, appear to be out of phase with local density increases. At first glance, this appears contrary to the positive correlation predicted by Immel et al. [2006] of nonmigrating tides that are responsible for launching upward plasma drifts that subsequently produce off-equatorial plasma density enhancements. Recall, however, that time constants for the plasma accumulation at higher altitudes and for off-equatorial dynamics must be taken into account before definitive comparisons of the variations of the ambient electric field and plasma density can attain meaningful status.

[39] On the nightside, VEFI data show an unexpectedly high degree of horizontal structuring including large (>1000 km) and small (<100 km) scales. Such structures appear in almost every nightside pass, even when plasma densities manifest only small to modest perturbations. Due to the absence of the PRE, evening sector irregularities in plasma density were generally weak to nonexistent. The type of electric field structuring shown in Figures 1b and 1c does not appear to be associated with typical spread F plasma depletions. Rather, the electric field structures are probably polarization effects of small-amplitude density perturbations in the F layer below the spacecraft. Causative plasma density structures may in turn be associated with neutral structures, such as those associated with gravity wave seeding from below the ionosphere, such as those observed in the DE-2 satellite data reported by Earle et al. [2008]. To maintain the continuity of eastward gravity-driven currents, positive/negative polarization charges accumulate on the negative/positive gradient sides of the density structures. Although small signs of density perturbations reach topside altitudes, polarization electric fields map along the magnetic field at roughly the local Alfvén speed and are readily observed in the C/NOFS data.

[40] On the other hand, Figure 1a shows a sequence of deep plasma depletions with strong eastward polarization electric fields within depleted flux tubes. On 14 June 2008, C/NOFS observed the formation and evolution of similar plasma density and electric field structuring in the postmidnight ionosphere [Burke et al., 2009]. Over the next 3 orbits individual depletions coalesced into a longitudinally broad (>10°) trench in which the plasma density fell below 10 cm−3. In this case, depletion formation was attributed to an eastward overshielding electric field associated with the passage of a high-speed stream in the solar wind. However, in the 24 h preceding the event in Figure 1a, the solar wind and interplanetary magnetic field were quiescent. We have followed the depletions observed in Figure 1a and found that they did not evolve but simply corotated to the dawn meridian. This phenomenology is not uncommon in the C/NOFS database and raises questions as to how such plasma depletions can form. Retterer [2009] describes simulations in which the ambient electric field ranged from weakly westward to zero. Due to the low altitude of the F layer during solar minimum, recombination produced very sharp positive gradients on the bottomside. In this case, the gravity-driven current caused small-amplitude irregularity seeds to grow over several hours into the nonlinear stage, punch through the F peak, and then rapidly rise in the topside.

5. Summary

[41] The main results revealed by the C/NOFS electric field and plasma density data shown here include the following.

[42] 1. DC electric field data include continuous vector measurements with meridional and zonal components each gathered with high time resolution. The data shown here were gathered on a unique satellite orbit (401 km by 867 km, 13° inclination) during a unique period of extreme solar minimum conditions.

[43] 2. The observed vector E × B drifts are organized into diurnal patterns that resemble previous results but deviate considerably with longitude.

[44] 3. The afternoon meridional E × B drifts are generally weak and even downward for some longitudes.

[45] 4. No prereversal, zonal electric field enhancement was observed at essentially any longitude. This is in agreement with previous satellite and radar observations during the June solstice season and weak solar activity conditions. Since C/NOFS does not journey below 400 km, it does not sample the upper atmosphere region where the neutral winds, and their subsequent polarization electric fields, are believed to be strongest. If such electric fields were produced at latitudes off of the magnetic equator, however, they would map along magnetic field lines to the higher altitudes encountered by C/NOFS.

[46] 5. Enhanced electric fields are observed at sunrise but only at longitudes between roughly 200° and 290°. These are isolated peaks in the meridional (upward) flow that represent a flow continuity of the eastward drifts that had persisted for a longer portion of presunrise hours and end abruptly at these times. Such enhanced sunrise meridional drifts were reported in the San Marco electric field data by Aggson et al. [1995] as well as in the ROCSAT ion drift data reported by Fejer et al. [2008]. The region of enhanced peaks covers the longitudes where the magnetic declination is significantly eastward, which is to be compared with the westward tilt of the terminator.

[47] 6. Nighttime electric fields are highly structured in both zonal and meridional components, even where the plasma density shows only modest variations. This structure is typically dominated by horizontal scale lengths of 500–1500 km along the satellite path and is observed at all altitudes sampled by the C/NOFS satellite (401–867 km).

[48] 7. The plasma density is observed along the C/NOFS orbit to be consistently organized by longitude, in both day and night conditions and at all altitudes.

[49] 8. The plasma drifts are observed to also be consistently organized by longitude, particularly during the day and at all altitudes. The plasma drift and plasma density variations with longitude are not in phase.

[50] 9. Zonal drifts at night are much weaker than those observed previously during solar maximum conditions. Nighttime zonal drifts after midnight are observed to be much weaker than those observed premidnight, possible due to the presumed reduced neutral winds in the postmidnight sector.

[51] As a result of the high duty cycle of C/NOFS measurements, vector DC electric field data can now be analyzed in detail both within an orbit and between consecutive orbits on a routine basis. Relationships between electric field and plasma density variations can be analyzed as functions of longitude, latitude, and local time. The data presented here clearly show that the electric fields and plasma density may vary greatly from orbit to orbit, yet maintain a persistent ordering, primarily by local time and longitude. C/NOFS has opened a window on a parametric range for the equatorial ionosphere that will draw our attention for years into the future.

Acknowledgments

[52] The Communication/Navigation Outage Forecast System (C/NOFS) mission, conceived and developed by the Air Force Research Laboratory, is sponsored and executed by the USAF Space Test Program.

[53] Robert Lysak thanks the reviewers for their assistance in evaluating this paper.

Ancillary